A New Approach to Plan-Space Explanation: Analyzing Plan-Property Dependencies in Oversubscription Planning
Authors: Rebecca Eifler, Michael Cashmore, Jörg Hoffmann, Daniele Magazzeni, Marcel Steinmetz9818-9826
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We run comprehensive experiments on adapted IPC benchmarks, and find that the suggested analyses are reasonably feasible at the global level, and become significantly more effective at the local level. We run comprehensive experiments on IPC benchmarks adapted to OSP as in previous work (Domshlak and Mirkis 2015; Katz et al. 2019), and on a collection of benchmarks we extended with action-set properties. |
| Researcher Affiliation | Academia | 1Saarland University, Saarland Informatics Campus, Saarbr ucken, Germany, 2University of Strathclyde, Computer and Information Sciences, Glasgow, UK, 3King s College London, Department of Informatics, London, UK |
| Pseudocode | No | No pseudocode or clearly labeled algorithm blocks are present in the paper. |
| Open Source Code | No | We implemented our approach in Fast Downward (FD) (Helmert 2006). (This refers to a third-party tool, not the authors’ own code release for their method). The paper does not provide a direct link or explicit statement about releasing their specific implementation code. |
| Open Datasets | Yes | We run comprehensive experiments on IPC benchmarks adapted to OSP as in previous work (Domshlak and Mirkis 2015; Katz et al. 2019), and on a collection of benchmarks we extended with action-set properties. for every IPC benchmark task (V, A, c, I, G) with smallest known plan cost C as per planning.domains (Muise 2016). |
| Dataset Splits | No | The paper describes the generation of OSP tasks from IPC benchmarks and the evaluation metrics, but it does not specify explicit training, validation, or test dataset splits. |
| Hardware Specification | Yes | The experiments were run on a cluster of Intel E52660 machines running at 2.20 GHz, with time (memory) cut-offs of 30 minutes (4 GB). |
| Software Dependencies | No | We implemented our approach in Fast Downward (FD) (Helmert 2006). (This mentions a software tool but does not provide its version number or any other software dependencies with specific versions). |
| Experiment Setup | Yes | The base planner called by our Sys S and Sys W algorithms on each search node runs forward search using h FF (Hoffmann and Nebel 2001), optionally with conjunction or trap learning. for every IPC benchmark task (V, A, c, I, G) with smallest known plan cost C as per planning.domains (Muise 2016), we obtained three OSP tasks by setting the cost bound to b = x C where x {0.25, 0.5, 0.75}. We used soft goals only, i. e., Gsoft = G and Ghard = . |